영상처리기법을 활용한 해상원유유출 범위 추정

Abstract

Oil spill accidents at sea not only destroy the ocean environment and neighboring ecosystem but also do harm people by the maleficent gas generated by the phase change of crude oil. Therefore, it is necessary to minimize the damage by initiating a quick response as soon as an oil spill occurs. Previous studies used the radar (SAR) image to calculate the oil spill area and determine the exact amount of oil spilt. SAR images, however, are difficult to access and therefore cannot be utilized in time. This paper introduce a method to estimate the oil spill area from the image obtained using various digital devices. Images containing oil spill scenes are analyzed using an image processing technique. The oil spill images used in the analysis are normally classified into two regions for the water and the oil. Those two regions are distinguished by the color difference based on the hue property of the HSV color space. A preprocessing approach that minimizes the errors possibly caused by the light reflection on the image is proposed. The histogram of the preprocessed image is analyzed and binarized to determine the regions of the water and the oil. When the binarization by the color cannot be accomplished, additional properties of an image such as saturation or brightness are used to estimate the oil spill area. A graphical user interface system is developed to help the user to estimate the ratio of oil spill region from an image.